Not sure what the top is, but I totally expect people to eventually get a few setups.
Modest, lower power, 80 to 100Mhz. Actually, for TV type signals, 108Mhz would divide right down to 13.5. That's suitable for PAL / NTSC.
Textbook. Sort of like Demoboard, 80Mhz 5Mhz Xtal was, only for the P2, that's probably 180-200Mhz. Maybe 250? This is the general target, like Demoboard was.
Both of those can likely be passive cooling, no extras required. Just don't wrap it in a blanket and run full tilt!
I am gonna run my eval pretty much standard. Gotta get through testing.
The beast Clocked way up, active cooling. Could be 400Mhz! Whatever it is that can be proven stable.
When we've got production chips, I'm going to build a beast mode one. Existing Intel overclock gear should work just fine.
Likewise I am going to run the P2 as is in the Engineering Sample board. We need to find a baseline for what could be expected for this chip. Clock speed, temperature, and a whole bunch of things. Clouding it up with fans, coolers, and who knows what else skews the data & the reason we are testing this new P2 chip in the first place. I certainly want to know the typical and absolute ratings of this chip. If that means a few of the P2 ES chips fail for whatever reason I certainly want to know that. When I was working for the defense industry, we had what was called Golden Samples, which we put thru electrical, thermal, and radiation cycling. After that, we put the Samples on a key chain and we engineers walked around for a month exposing them to shock, electrostatic discharge, and what ever else was in our pockets. They got beat up pretty bad, some failed, but we learned a lot.
Sort of like the 5 year old kid one for software. Slide the machine over to a "test subject" and wait... If it's gonna go south on unexpected input, that's one quick way to find out!
I agree with you guys about these first chips and boards. I'm going to be using my P2-EV AS IS with no mods. There is enough to be learning about what does and doesn't work without clouding the issue with unknowns or frying one of these unique boards.
I've been waiting to build a P2 board for thirteen years now. I'm getting to a point where I probably won't be doing it too many more times, I want to get this one right.
Soon enough the P2 will be available in hundreds or thousand piece quantities. ( Probably by the time I get this layout done) I'm looking forward to the first time I manage to get an LED to blink with a P2 board I designed and build myself, but eventually the "New" will wear off and it will be time to "Turn and Burn". I'm trying to do this design so that it has the capability of doing the maximum amount of "turning" and a minimum amount of "Burning". I have no intention or likely need to use the maximum ability of the P2 as is, as designed, but SOMEONE is going to want/need to push the very edge. If it takes a chiller or two to accomplish that (and I can fit them in) I'd like to have those things in the design from the very start instead of having to redo the design later.
I finished the first layout of my P2 industrial board. Attached are some photos. Some of you mentioned how industrial designs need heavy duty connections. The relay board shown attached is what I had in mind for connecting to industrial systems. This one uses the connectors on the left to mount Opto relays. The connectors on the bottom and right are still open for high-speed stuff, analog I/O etc. . I can see the main P2 board mounted on a much larger board with all the necessary filtering power connections, etc. for major control designs, but smaller add-on's for smaller control apps.
I already have some simple stepper drive boards designed to mount on the P2 main board ( I'm calling it the INDY ). The next thing I want to work on will be some analog boards for +-10 volt DC motor drives, current sensing, temperature control, etc. Any suggestions on modules that might be useful? I'll have to wait for the next P2 silicon to have encoder input working with the smart pins but I'm hoping to have something working with my Eval board if I can adapt one of the P1 encoder functions from the OBEX. Is anyone else working on motion control who might be interested?
I've seen some recent activity on the forum about boards to make use of the new chips coming soon. I thought I'd bring up my "Indy" (industrial ) design again.
I haven't had any of these boards manufactured yet but with "Real" chips almost available I thought I'd test the waters to see if any one else might be interested in ordering bare boards ( or possibly a "Kit" with the basic components and connectors.)
I designed this board with hand solderable components (805) and connectors sturdy enough to use in the "Real" world.
I still need to make a couple of changes ( holes below the P2 big enough for a solder tip, etc. ) But I'll probably try to finalize the design over the next few days and get my first boards ordered.
I've also been considering doing a re-design that mounts a Raspberry Pi "Front End" board.
The INDY design certainly looks robust. Having a Rasperry Pi interface is something to consider. BUT with all the new digital and analog I/O a new buss standard is certainly something you want to do.
One thing I am most concerned about is having a robust analog input protection circuit, one that protects the P2 "analog" pin(s) and has a reasonable dynamic voltage and frequency (many MHz) range
Likewise for the P2 analog output pin(s). Something to drive a light, motor, or power amplifier with the same decent dynamic range (many MHz) at various (TBD) voltages and currents. I don't know if you have already designed this, but it would be a great help help for us that don't have your background in industrial controller designs.
I've seen some recent activity on the forum about boards to make use of the new chips coming soon. I thought I'd bring up my "Indy" (industrial ) design again.
I haven't had any of these boards manufactured yet but with "Real" chips almost available I thought I'd test the waters to see if any one else might be interested in ordering bare boards ( or possibly a "Kit" with the basic components and connectors.)
I designed this board with hand solderable components (805) and connectors sturdy enough to use in the "Real" world.
I still need to make a couple of changes ( holes below the P2 big enough for a solder tip, etc. ) But I'll probably try to finalize the design over the next few days and get my first boards ordered.
I've also been considering doing a re-design that mounts a Raspberry Pi "Front End" board.
Anyone interested?
Definitely interested in what you already have. Don't need the Pi necessarily
I've also been considering doing a re-design that mounts a Raspberry Pi "Front End" board.
Attached is an image of how I did a Pi-pinout variant of P2D2. This uses 4 solder bridge options for PiGNDs or P2 IOs, and I'd expect most users to default to the P2 superset initially, and just skip the Pi extra redundant GNDs. Think of that as a superset standard that supports 32 IO and allows Pi-Board connections, and even Pi replacement.
However, to be useful, you do not have to use all Pi IO pins, there are displays that start with as low as 6, like this new OLED one (128x64) - it uses just 6 pins for i2c
some other Pi boards use a split header, to get i2c one end, then a gap to SPI pins, and I've seen claims of 125MHz Pi SPI .
This LCD series uses a 26 pi header and claims 125MHz SPI
That's a neat idea Jmg! however... what I was considering was making a P2 board that the PI itself would mount to. There's a lot of stuff on the $35 Raspberry Pi 4: Ethernet, USB 3.0, Wifi , cameras, HDMI, etc that would cost a lot to put on a P2 board... so let the Pi be the "front end" or operator interface for the P2. The Pi could do the operator interface stuff with the P2 doing the heavy lifting for process control/audio/robotics etc.
I haven't looked into the specs yet, but the Pi already has voltage regulation for 5 and 3.3 volts, it's probably fairly robust and there are LOTS of things that are needed in engineering/industrial control/laboratory that won't run on the P2.
I use Labview to control some of my Medical test equipment going through Labjack I/O systems. Honestly, it's a pretty lousy way to go because Windows screws up some critical timing, however, I needed analog I/O so I wound up with this combination.
I did a quick serial interface between the Propeller and Labview a few years back, but at the time it just wasn't cost effective to turn the Prop into another "Labjack". The P2 on the other hand would make a pretty awesome REAL TIME controller ( they cost close to $3000 from National Instruments) .
For some of my manufacturing machines I control with the P1, I just reprogram the Prop to change manufacturing parameters. ( times, temps, cycles, etc.) I wrote a "Windows like" front end for the prop years ago, but honestly, as a programmer, changing a few numbers and pressing F10 was less hassle to use. ( not so for most people accessing process control equipment however )
Now, with the P2 "Real" I still think it will end up easier and probably cheaper having some sort of machine (like the lap-top I am writing this on) do what it does best and using the P2 for the REAL TIME things It does best.
I didn't realize it until a few weeks ago, but my Labview development environment has a simple little function that automatically turns your "Instrument" (or control front end ) into a web page that you can access from most web browsers. It should be simple to set up the P2 doing just about any control/process monitoring function (security, monitoring, etc) and have it all be accessible from anywhere on the planet. In short... a $35 Pi + P2 + a bit of industrial I/O filtering ($150 - $250 range) SHOULD let us do things that cost over $3000 to do now.
I suspect there are other easy web-access pathways available on the Pi that we could use other than Labview... but this is "The devil I know".
I DO like the idea of using other standard modules that are already out there in common use. Anyone got any suggestions for a standard that might be an easy add for the INDY board?
That's a neat idea Jmg! however... what I was considering was making a P2 board that the PI itself would mount to. There's a lot of stuff on the $35 Raspberry Pi 4: Ethernet, USB 3.0, Wifi , cameras, HDMI, etc that would cost a lot to put on a P2 board... so let the Pi be the "front end" or operator interface for the P2. The Pi could do the operator interface stuff with the P2 doing the heavy lifting for process control/audio/robotics etc...
Yup, because the P2 pins are completely soft, you can have the P2 act as either host or slave.
A P2D2Pi can slave to a Pi, if you want to use Pi as the HMI, or if the Pi was doing modest work, it can be replaced by the P2. Many options....
There are 2 x 40 pin headers on P2D2Pi, so you can even connect 2 Pi's with some mechanical gymnastics...
For the type of things I have in mind, I can't imagine needing more than one Pi connected to a P2... but to each his own. I've always wanted to build a Beowulf cluster, mainly just to see if I could do it. I have no weather systems to model or such.
If I could give my flight simulator a boost it might be worthwhile but a new graphics card would probably be quite a bit easier.
I suspect the P2 WILL wind up being used for such things as connecting multiple Pi's together into super-computer nodes. I look forward to seeing what this child of such difficult birth will become as it grows up.
Comments
Modest, lower power, 80 to 100Mhz. Actually, for TV type signals, 108Mhz would divide right down to 13.5. That's suitable for PAL / NTSC.
Textbook. Sort of like Demoboard, 80Mhz 5Mhz Xtal was, only for the P2, that's probably 180-200Mhz. Maybe 250? This is the general target, like Demoboard was.
Both of those can likely be passive cooling, no extras required. Just don't wrap it in a blanket and run full tilt!
I am gonna run my eval pretty much standard. Gotta get through testing.
The beast
When we've got production chips, I'm going to build a beast mode one. Existing Intel overclock gear should work just fine.
Sort of like the 5 year old kid one for software. Slide the machine over to a "test subject" and wait... If it's gonna go south on unexpected input, that's one quick way to find out!
I've been waiting to build a P2 board for thirteen years now. I'm getting to a point where I probably won't be doing it too many more times, I want to get this one right.
Soon enough the P2 will be available in hundreds or thousand piece quantities. ( Probably by the time I get this layout done) I'm looking forward to the first time I manage to get an LED to blink with a P2 board I designed and build myself, but eventually the "New" will wear off and it will be time to "Turn and Burn". I'm trying to do this design so that it has the capability of doing the maximum amount of "turning" and a minimum amount of "Burning". I have no intention or likely need to use the maximum ability of the P2 as is, as designed, but SOMEONE is going to want/need to push the very edge. If it takes a chiller or two to accomplish that (and I can fit them in) I'd like to have those things in the design from the very start instead of having to redo the design later.
I already have some simple stepper drive boards designed to mount on the P2 main board ( I'm calling it the INDY ). The next thing I want to work on will be some analog boards for +-10 volt DC motor drives, current sensing, temperature control, etc. Any suggestions on modules that might be useful? I'll have to wait for the next P2 silicon to have encoder input working with the smart pins but I'm hoping to have something working with my Eval board if I can adapt one of the P1 encoder functions from the OBEX. Is anyone else working on motion control who might be interested?
I haven't had any of these boards manufactured yet but with "Real" chips almost available I thought I'd test the waters to see if any one else might be interested in ordering bare boards ( or possibly a "Kit" with the basic components and connectors.)
I designed this board with hand solderable components (805) and connectors sturdy enough to use in the "Real" world.
I still need to make a couple of changes ( holes below the P2 big enough for a solder tip, etc. ) But I'll probably try to finalize the design over the next few days and get my first boards ordered.
I've also been considering doing a re-design that mounts a Raspberry Pi "Front End" board.
Anyone interested?
One thing I am most concerned about is having a robust analog input protection circuit, one that protects the P2 "analog" pin(s) and has a reasonable dynamic voltage and frequency (many MHz) range
Likewise for the P2 analog output pin(s). Something to drive a light, motor, or power amplifier with the same decent dynamic range (many MHz) at various (TBD) voltages and currents. I don't know if you have already designed this, but it would be a great help help for us that don't have your background in industrial controller designs.
Definitely interested in what you already have. Don't need the Pi necessarily
Attached is an image of how I did a Pi-pinout variant of P2D2. This uses 4 solder bridge options for PiGNDs or P2 IOs, and I'd expect most users to default to the P2 superset initially, and just skip the Pi extra redundant GNDs.
Think of that as a superset standard that supports 32 IO and allows Pi-Board connections, and even Pi replacement.
However, to be useful, you do not have to use all Pi IO pins, there are displays that start with as low as 6, like this new OLED one (128x64) - it uses just 6 pins for i2c
some other Pi boards use a split header, to get i2c one end, then a gap to SPI pins, and I've seen claims of 125MHz Pi SPI .
This LCD series uses a 26 pi header and claims 125MHz SPI
P2D2Pi pinout mapping :
I haven't looked into the specs yet, but the Pi already has voltage regulation for 5 and 3.3 volts, it's probably fairly robust and there are LOTS of things that are needed in engineering/industrial control/laboratory that won't run on the P2.
I use Labview to control some of my Medical test equipment going through Labjack I/O systems. Honestly, it's a pretty lousy way to go because Windows screws up some critical timing, however, I needed analog I/O so I wound up with this combination.
I did a quick serial interface between the Propeller and Labview a few years back, but at the time it just wasn't cost effective to turn the Prop into another "Labjack". The P2 on the other hand would make a pretty awesome REAL TIME controller ( they cost close to $3000 from National Instruments) .
For some of my manufacturing machines I control with the P1, I just reprogram the Prop to change manufacturing parameters. ( times, temps, cycles, etc.) I wrote a "Windows like" front end for the prop years ago, but honestly, as a programmer, changing a few numbers and pressing F10 was less hassle to use. ( not so for most people accessing process control equipment however )
Now, with the P2 "Real" I still think it will end up easier and probably cheaper having some sort of machine (like the lap-top I am writing this on) do what it does best and using the P2 for the REAL TIME things It does best.
I didn't realize it until a few weeks ago, but my Labview development environment has a simple little function that automatically turns your "Instrument" (or control front end ) into a web page that you can access from most web browsers. It should be simple to set up the P2 doing just about any control/process monitoring function (security, monitoring, etc) and have it all be accessible from anywhere on the planet. In short... a $35 Pi + P2 + a bit of industrial I/O filtering ($150 - $250 range) SHOULD let us do things that cost over $3000 to do now.
I suspect there are other easy web-access pathways available on the Pi that we could use other than Labview... but this is "The devil I know".
I DO like the idea of using other standard modules that are already out there in common use. Anyone got any suggestions for a standard that might be an easy add for the INDY board?
Yup, because the P2 pins are completely soft, you can have the P2 act as either host or slave.
A P2D2Pi can slave to a Pi, if you want to use Pi as the HMI, or if the Pi was doing modest work, it can be replaced by the P2. Many options....
There are 2 x 40 pin headers on P2D2Pi, so you can even connect 2 Pi's with some mechanical gymnastics...
If I could give my flight simulator a boost it might be worthwhile but a new graphics card would probably be quite a bit easier.
I suspect the P2 WILL wind up being used for such things as connecting multiple Pi's together into super-computer nodes. I look forward to seeing what this child of such difficult birth will become as it grows up.